Atomic Force Microscopy (AFM) is a powerful tool for nano-size imaging. The advantage of AFM is that it can get extraordinary high resolution image at atom level. However, AFM obtains the sample topography image through scanning on the top of sample line by line, therefore it takes couples minutes to get an image and this negative point makes it difficult to continuously observe surface change during manipulation. In this paper, a novel approach for compressive sensing based video rate AFM imaging system is proposed. In this method, compressive sensing is used for sampling topography information of sample surface efficiently. Compressive sensing could use fewer measurements for data sensing to recovery the image through image reconstruction algorithm. This technique decreases the scanning time for AFM scanner because of fewer measurements needed. The video rate for this new approach could reach as high as 1.75 seconds per frame.